Alzheimer's disease (AD) is among the most devastating conditions that affect older Veterans today, and despite decades of intense research, there is still no effective treatment. Convincing epidemiologic evidence suggests that cancer survivors have a lower risk of AD, and that people with AD have a lower risk of cancer. In this proposal, we pursue the hypothesis that this inverse association may hold the key to new approaches to prevention and treatment of AD. Our pilot data in a national cohort of Veterans shows that survivors of many cancer types have a decreased risk of AD that is independent of cancer treatment, and that those who receive chemotherapy have an even lower risk. The protective effect of chemotherapy did not seem to be explained by earlier death or less frequent diagnosis of other conditions among cancer patients. It is quite plausible that anti-cancer treatments could decrease AD risk, since some common drugs are known to improve outcomes in mouse models of AD through multiple mechanisms including stabilization of microtubules and dissolution of tangles. Other cancer drugs interrupt the cell cycle in its early stages, and may thus prevent neuronal cell death. Furthermore, cancer and AD share many key pathophysiologic features, including oxidative stress, metabolic dysregulation, DNA damage, and inflammation. Agents that suppress these pathways might be used as chemoprevention for both diseases. One example is the diabetes drug metformin, for which there is emerging evidence of both anti-neoplastic and neuroprotective effects. In this proposal, our multidisciplinary research team will further explore the epidemiologic and biological link between AD and cancer using the rich resources of the Massachusetts Veterans Epidemiology Research and Information Center and the Department of Statistical Genetics at the Harvard School of Public Health. In Aim 1 we will investigate the association between 20 different cancer types and AD in our national dataset of over 3.5 million Veterans. We will refine our pilot analyses by validating a more accurate definition of AD, confirming the cancer diagnosis through the VA Cancer Registry, and requesting linked Medicare data. In Aim 2, we will examine the relationship between particular classes of anti-neoplastic therapy and the risk of AD in both the national VA dataset and our more detailed pharmaco-epidemiology database. We will also determine whether regular users of metformin have a lower risk of AD. In Aim 3, we will perform genome-wide meta- analyses of published studies on cancer and AD to identify shared genetic variants of both diseases. We will gain insight into common biological pathways using novel techniques such as gene-set enrichment, pathway and network analysis. Successful completion of these studies will generate important hypotheses about drugs that could be repositioned as treatment or chemoprevention for AD, and lead to clinical trials that would directly benefit US Veterans. Characterization of the biological overlap between cancer and AD will likely yield new pathophysiologic insights and could identify new targets for therapy.